Implementation of a Hierarchical Model Predictive Control Strategy for Coordinated Energy Management

Abstract

Multi-Energy Systems (MES) require coordinated control strategies capable of handling the heterogeneous dynamics of different energy domains. This work presents the implementation and experimental validation of a hierarchical Model Predictive Control (MPC) architecture applied to a real multi-energy facility. The proposed approach uses a centralized high-level MPC to optimize energy exchanges and storage management, while decentralized local MPC controllers regulate subsystem dynamics on separate virtual machines using MQTT (Message Queuing Telemetry Transport)-based communication. Experiments conducted on the RSE’s Distributed Energy Resources Test Facility (DER-TF) show that the hierarchical strategy significantly improves overall system performance compared to independent local control. In particular, coordinated operation enables more effective use of Combined Heat and Power (CHP) generator and thermal storage, achieving up to 24% reduction in operating costs over a six-hour test window. These results demonstrate the practical feasibility of hierarchical MPC for real-world multi-energy coordination and its potential to enhance efficiency in integrated energy systems.